Face Landmark Tracking for Accurate Contactless Biometric Measurement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing non-contact biometric measurement methods struggle with movement noise and lighting changes, leading to reduced accuracy and inefficiency, particularly in daily life applications and medical settings, and the process of obtaining biometric information at medical institutions is cumbersome.
Innovation Solution
A method that includes detecting face landmarks, tracking minimal movement sections, extracting skin area pixel change data, and using a learning model to measure biometric information robustly, with predefined measurement spot areas for high accuracy, and integrating user authentication for simplified medical procedures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If image-based non-contact measurement method is used, then convenience and real-time performance are improved, but measurement accuracy deteriorates due to movement noise and lighting changes
Solution Approach 1:
The patent segments the face area into multiple regions of interest (eyes, nose, mouth, cheeks) and processes each region separately to extract blood flow information. This segmentation allows the system to focus on specific areas less affected by movement noise and lighting changes, thereby maintaining measurement accuracy while preserving the convenience of non-contact measurement.
Solution Approach 2:
The patent introduces an intermediary processing step that uses face landmark detection and region-of-interest segmentation as a mediator between the raw image data and the final biometric measurement. This intermediary process filters out movement noise and lighting variations before extracting the actual blood flow signals, resolving the contradiction between convenience and accuracy.
2Device complexity
If standard heart rate data with relative power density clustering is used, then heart rate determination is simplified, but responsiveness to face movement deteriorates
Solution Approach 1:
The patent implements a dynamic approach by continuously tracking face landmarks and adaptively adjusting the regions of interest based on detected face movements. Instead of using fixed clustering thresholds, the system dynamically modifies the analysis regions to follow facial features, maintaining adaptability to movement while keeping the overall process relatively simple through automated tracking.
Solution Approach 2:
The patent incorporates feedback mechanisms where the detected face movement information is fed back into the region selection and signal processing stages. This feedback loop allows the system to automatically adjust to face movements in real-time, improving responsiveness while maintaining process simplicity through closed-loop control.
3Measurement precision
If multiple processing steps including access check, identity check, and biometric measurement are implemented, then measurement accuracy is improved, but processing time and complexity increase
Solution Approach 1:
The patent merges the face recognition and biometric measurement processes into a single integrated workflow. By using the same captured image and detected face landmarks for both identity verification and blood flow analysis, the system eliminates redundant processing steps and reduces overall processing time while maintaining measurement accuracy through the unified approach.
Solution Approach 2:
The patent implements a multi-functional system where the face detection and image processing infrastructure serves multiple purposes: identity authentication, region-of-interest determination, and biometric measurement. This universality allows a single processing pipeline to accomplish multiple tasks that would traditionally require separate steps, reducing time loss while preserving accuracy.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Accurately measures biometric information despite user movement and lighting changes, enabling quick and efficient biometric data capture for telemedicine and streamlined medical institution access.
Implementation Method 1
photographing an image including a user's face
Implementation Method 2
detecting spectroscopic fine changes in the face area using a camera
Data Source
AI summary
A method for measuring a non-contact biometric information, includes steps of photographing an image including a user's face; detecting a face area in a frame of a photographed image; detecting a landmark within the face area and starting a tracking mode when a specific landmark is detected; detecting a movement of the face area in the tracking mode and determining a section in which the movement of the face is less than a reference value; and extracting a skin area within the face area in the determined section and acquiring pixel change component data of the skin area.